The invention will be described more particularly in relation to its applications to the characterization of ultrashort laser pulses having space-time couplings, even though these uses are not exclusive. “Ultrashort laser pulse” is taken to mean a picosecond pulse, that is to say of duration comprised between around 0.1 and 100 ps, or a femtosecond pulse, that is to say of duration less than or equal to 100 fs=0.1 ps. The durations extend to mid-height of the intensity profile. These pulses have a relative wide spectral band, that is to say typically of the order of several tens of nanometers, or even of the order of several hundreds of nanometers.
Ultrashort laser pulses have numerous scientific and technological applications; they can be amplified up to energies of several joules and form beams, referred to as “pulsed beams”, the diameter of which goes from several millimeters to several centimeters as a function, notably, of their power.
Generally speaking, the temporal properties of the electromagnetic field of a pulsed beam can vary spatially or, in an equivalent manner, the spatial properties of the electromagnetic field of a pulsed beam may be time dependent. For example, the pulse duration may depend on the position (x,y) in the beam. In the present description, unless stated otherwise, a beam propagating in a direction “z” will be considered, the “x”, “y” and “z” axes forming an orthonormal coordinate system.
When such a dependency exists, the field E(x,y,t) cannot be expressed in the form:E(x,y,t)=E1(t)×E2(x,y)where E1(t) is a time function and E2(x,y) a space function. It is then said that the beam has space-time coupling (STC).
Space-time couplings may lead notably to distortion of the intensity front of a pulsed beam, illustrated by means of FIGS. 1a and 1b. FIG. 1a illustrates the ideal case, according to which the electromagnetic energy of an ultrashort pulsed beam propagating in the direction z is spread out in a very thin disc of diameter D and of thickness cT, where c is the speed of light and T the duration of the pulse. In the example of FIG. 1a, D=8 cm and cT=10 μm, which corresponds to a pulse duration of around 33 fs. In order to maximize the light intensity obtained at the focus, which is generally desired, said disc must be as “flat” as possible. To characterize this spatial distribution of the energy, the expression “intensity front” of the laser is used. The notion of intensity front must not be confused with that of “wave front”.
In practice, and notably in the case of high power lasers with large beam diameter, the intensity fronts may not be flat but distorted, as illustrated in FIG. 1b. Consequently, the pulse peak may be reached at different instants in the different points of the section of the beam in the plane (x,y), and the pulse duration may also vary from one point to another. Other types of space-time couplings are also possible, such as for example a rotation of the wave fronts over time. Techniques for measuring these couplings have been proposed, but they remain limited in their performances, are complex to implement and are unsuited to large size beams, typically coming from high power sources. Consequently, these techniques are not widespread. In fact, the larger a light beam, the more likely it is to have space-time coupling. It is thus in particular for such light beams that it is important to be able to carry out a measurement of space-time couplings. These techniques are listed hereafter.                The “SPIDER-2D” technique, costly and complex, imposes a limitation of the size of the characterized beam. SPIDER-2D enables the reconstruction of the characterized beam as a function of time t and a transversal direction, x or y.        The “STRIPED FISH” technique is simpler to implement and cheaper than SPIDER-2D. STRIPED FISH moreover enables the reconstruction of the characterized beam as a function of time t and two transversal directions x and y. Nevertheless, STRIPED FISH also imposes a limitation of the size of the characterized beam, requires the use of a reference beam, which can prove very difficult to obtain, and only enables small spectral sampling.        The “HAMSTER” technique is described in the article of Cousin et al., “Three-dimensional spatiotemporal pulse characterization with an acousto-optic pulse shaper and a Hartmann-Shack wavefront sensor”, Optics Letters 37, 3291 (2012). HAMSTER uses an acoustic-optic modulator and a 2D wave front sensor of Shack-Hartmann type in order to carry out a time measurement at a point of the characterized beam, then to measure the spatial wave front of different spectral sections of the characterized beam. At the end of two series of measurements, HAMSTER arrives at a complete space-time reconstruction of the characterized beam, that is to say as a function of time t and two transversal directions x and y. The HAMSTER technique nevertheless involves a certain complexity and a high cost, on account notably of the use of an acoustic-optic modulator. On the other hand, the HAMSTER technique is unsuitable for large diameter beams.        The “SEA TADPOLE” technique consists in collecting light at different points of the characterized beam with a first optical fibre, while moving said first optical fibre to the different points of the beam. An auxiliary beam is injected into a second optical fibre. The output ends of the first and second optical fibres are placed close to each other, in such a way that in diverging, the beams leaving said first and second optical fibres overlap spatially and produce spatial interferences. These spatial interferences are spectrally resolved using a spectrometer, to obtain an interferogram. This interferogram makes it possible to determine the spectral phase between the beam injected into the first fibre, and that injected into the second fibre. The spectral properties of the light collected at a point of the characterized beam are thus compared with those of the auxiliary beam. By moving the first fibre to a plurality of points of the characterized beam, each of these points is compared with the auxiliary beam, which makes it possible to reconstruct the spectral phase of the characterized beam. Unlike the SPIDER-2D and STRIPED FISH techniques, the SEA TADPOLE technique has the advantage of not limiting the size of the characterized beam. In the same way as for STRIPED FISH, the mounting is relatively simple and inexpensive and the reconstruction of the characterized beam is carried out according to three dimensions (x, y, t). However, the characterization of a beam by SEA TADPOLE requires a large number of laser shots, the spectral phase being determined point by point. The necessity of carrying out several laser shots imposes that the laser beam to characterize is stable and reproducible shot-to-shot, which is not always the case for femtosecond lasers, and in particular for high power femtosecond lasers. Another limitation of the SEA TADPOLE technique is constituted by the use of optical fibres, which introduce random phase fluctuations,        The closest state of the art is constituted by the technique referred to as “MUFFIN”, which is described in the patent FR 2976663 (A1). The SPIDER-2D, STRIPED FISH and SEA TADPOLE techniques evoked previously are also described in a detailed manner in the preamble of patent FR 2976663(A1). The MUFFIN technique represents an improvement to the SEA TADPOLE technique. Instead of using two optical fibres—a first optical fibre that is moved successively to N points of the characterized beam and a second optical fibre that serves as reference, MUFFIN proposes using directly a set of N optical fibres. The input ends of these N fibres collect light at N different points of the characterized beam. The output ends of these N fibres are placed next to each other in a line, so that the beams on leaving said fibres overlap and spatially interfere with each other. The MUFFIN technique thus arrives, in a single shot and without necessarily having available an auxiliary beam, at the result of SEA TADPOLE. MUFFIN does not eliminate straight away the problem of phase fluctuation in the optical fibres, already identified for SEA TADPOLE. Such phase fluctuations may prevent complete reconstruction of the characterized beam. A solution to this problem has been proposed in the patent cited above, but it makes the use of the MUFFIN technique more complex. Furthermore, it is difficult with the MUFFIN technique to use a large number of optical fibres. In practice, the maximum number of optical fibres is of the order of several tens. Consequently, the MUFFIN technique only enables a limited spatial sampling of the characterized beam.        